The traditional fault diagnosis method of high-voltage circuit breaker based on artificial neural network has a shortcoming that the complex of network structure because of the high-voltage circuit breaker fault feature information is too large

and in the previously proposed self-organizing feature map network to verify the fault diagnosis also exists when the training process is time-consuming

slow diagnosis of the shortcomings.So

a rough sets theory combined with self-organizing feature map network has proposed in this paper

firstly

the attributes of the fault data are evaluated by the rough set theory and the minimum attribute set is found

the redundant attributes in the feature information are eliminated

the reduction decision table is obtained

the new fault characteristic data is formed

the new fault characteristic data is also as the input of self-organizing feature map.The new method ensure that the overall accuracy rate achieved 91%

and simplifying the network structure

improve the running time and diagnostic rate after verification

so this algorithm can be effectively implemented in fault diagnosis.